What is nudging anyway?
Since Richard Thaler and Cass Sunstein published their book ‘Nudge’ in 2008, behavioural science teams have popped up all over the world. They followed in the footsteps of the Behavioural Insights Team, the original ‘Nudge Unit’ founded 2010 by the psychologist David Halpern, who last month announced that he was stepping away from his day-to-day role in running BIT. Nudging has become popular in part because it offers the possibility of changing behaviour using only limited resources. But it also has a mystic founded on claims of a more realistic and deeper understanding of human behaviour and decision-making. This is contrasted with traditional economics and policy levers which are seen as having unrealistic and naive expectations about the way people think and behave.
In Nudge, Thaler and Sustein explicitly contrast their conception of Homo sapiens with economists’ Homo economicus, the latter concept assuming that people ‘can think like Albert Einstein, store as much memory as Google does in the cloud, and exercise the will power of Mahatma Gandi’. This is a characterisation of economic assumptions that is widely held and I’ve heard many behavioural scientists who don’t have a background in economics dismiss economic models on the basis that they make unrealistic assumptions about human rationality and capacities. But to understand what behavioural economics and nudging actually are, you need to understand the orthodox model of human behaviour within economics and why ‘rational’ in that context means something subtly different to how we use it in everyday language.
When we use ‘rational’ in our normal lives we mean someone making decisions that we think are sensible and align with objective facts. Thus Jeremy Clarkson buying an Alfa Romeo as part of a Grand Tour challenge, despite the fact that the car is objectively worse but more expensive than the alternatives, is seen as ‘irrational’ and sets up an entertaining TV segment where Richard Hammond and James May point out all of the car’s flaws. But from an economist’s point of view, Clarkson isn’t being irrational at all. The price of the Alfa Romeo reflects its objective qualities but also its design, style and the feelings people have for it (its ‘brand’). All those things are part of the ‘utility’ (benefits) the buyer gets from a product. From an economic perspective, Clarkson would only start to be irrational if he radically and randomly changed his mind about how much he valued or was prepared to pay for an Alfa Romeo from day to day. This is because within economics, the meaning of the term ‘rationality’ is closer to the way we use ‘consistent’ in everyday language than how we use ‘rational’.
Classical economics and rational choice theory use other assumptions that actually reflect a deep and profound understanding of human behaviour. These include diminishing marginal returns (the assumption that a second slice of pizza is worth more to you than the tenth), delay discounting (a precise way of calculating the degree to which a future reward is less valued than a current one), and ‘risk aversion’ (which means preferring a guaranteed £10 to a 50/50 bet where you could win either £2 or £20). These core concepts are used within a constrained optimisation framework to analyse people’s decisions, and far from being unrealistic, this approach provides an extraordinarily powerful model for predicting human behaviour. (Arguably its biggest flaw is not its ability to predict behaviour but its ability to explain it.)
However, economic models aren’t perfect and behavioural economics explores the ways that human behaviour departs from traditional assumptions in meaningful ways. Some of these are discussed in Nudge and include a range of biases, such as anchoring (people’s judgements being influenced by an arbitrary reference point), the availability heuristic (the assumption that what is most easily recalled is most important), the representativeness heuristic (being overly influenced by things such as stereotypes) and loss aversion (valuing the loss of something more than the gain of an equivalent amount). Behavioural economics also introduces other departures from traditional economic models, such as hyperbolic discounting, which is not irrational from an economics perspective but rather is a more sophisticated way of understanding how future rewards are discounted in the present.
Some of these biases and modifications to traditional economic theory make a meaningful difference to the predictions the models make, but often the difference is very small. A rough analogy might be the difference between classical mechanics and relativity. The difference is meaningful when objects are approaching the speed of light, but classical mechanics is accurate enough to get your spaceship to the moon and back. In other words, for most everyday applications, the difference between relativity and classical mechanics is so minor as to be irrelevant.
Sometimes ‘nudging’ is used as an informal term for ‘behavioural economics’ but Thaler and Sunstein give it its own definition, which is ‘any aspect of the choice architecture that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic incentives’. This definition was translated into the framework MINDSPACE by Halpern and others in pre-Nudge Unit days and later into EAST by the BIT team. MINDSPACE is an acronym that stands for messenger, incentives, norms, defaults, salience, priming, affect, commitments, and ego, while EAST boils this list down into easy, attractive, social and timely. These frameworks encourage policy-makers to focus on the ways in which choices are presented to people, and to think about how the choice architecture can be altered to influence the decisions people make.
While this might seem clear, what is less clear is whether nudges in fact leverage the ‘deeper insights’ of behavioural economics. Take the first of the four EAST principles, ‘easy’. Easy can be applied through creating defaults, and an example of this given in EAST is the policy of automatically enrolling employees into pension schemes rather than requiring them to opt in, which is reported to have resulted in participation rates in large firms rising from 61% to 83%. This is undoubtedly a positive change in behaviour and implies that people were not actively declining to enrol in pension schemes. It’s not clear, though, that the lack of participation was due to ‘irrationality’ in the economic sense.
Before the default enrolment scheme was introduced, signing up to a pension could be complex and involve trying to make choices based on limited information and understanding, which is very stressful. These things represent immediate costs which are weighed against benefits that won’t be realised for a long time, which means that it is not surprising the people might put off the decision to some unspecified time in the future. This behaviour can be explained by the classical economics concept of delay discounting rather than the behavioural economics one of hyperbolic discounting. Hyperbolic discounting explains the more specific behaviour of deciding to do something (or not do something) at a future date but then reversing that decision when the time comes (eg deciding that you’re not going to drink when you get to a party but giving in to the temptation when you’re actually there). This may have been happening in the case of pension decisions - people may have been fixing a date when they would sort out their pension but deciding not to bother when that date arrived. But the concept is not needed to understand why making things easy helps ‘nudge’ people in certain directions, particularly when the costs and benefits are more finely balanced.
A similar analysis can be applied to the other principles in MINDSPACE and EAST. For example the benefit of making something attractive (the A in EAST) does not need to leverage ‘biases’ or address ‘irrationality’ because utility is not just based on financial incentives. Coke and Pepsi spend billions of dollars making their products attractive and the fact that this influences consumers decisions is not because consumers are being ‘irrational’ it’s because the drinks have a symbolic, signalling effect that provides utility alongside their flavour or thirst-quenching properties.
Nudging therefore is not the same as behavioural economics and does not solely or perhaps even mostly rest on modifications to classical economics principles. It is predominantly an approach that directs policy-makers and those designing interventions to closely examine the environment within which decisions are made and to consider how small changes to those environments can have meaningful effects on behaviour. This is both a strength and a limitation. The strength is that it addresses the problem that policy-makers have often failed to truly consider or take into account the context within which policies operate. The limitation is that it can encourage them to avoid considering the core underlying motivations or financial incentives that are the major levers of change.